搜索“Agent”

zon benchmark built to measure learning from real-world environments . The benchmark comprises 134 realistic and diverse tasks spanning six major capability domains, each allowing agents to operate continuously for at least 12 hours. Based on long-horizon interaction runs across these tasks, we find that the overall agent performance in "environment learning" closely follows a log-sigmoid curve , with an average goodness of fit of R² = 0.998. Across frontier models from different generations, ag

匹配字段正文

rts Weigh the Risks, Rules and Race Ahead Shana Lynch Sep 22, 2026 News In a new series called Prompt Response, Stanford’s Surya Ganguli, Diyi Yang and Rob Reich examined emergent agent behavior, recursive self-improvement, independent evaluation and whether a kill switch can make advanced AI safer. In a new series called Prompt Response, Stanford’s Surya Ganguli, Diyi Yang and Rob Reich examined emergent agent behavior, recursive self-improvement, independent evaluation and whether a kill switc

匹配字段正文

ew Stanford research finds they often don't measure what they claim to. In a new series called Prompt Response, Stanford’s Surya Ganguli, Diyi Yang and Rob Reich examined emergent agent behavior, recursive self-improvement, independent evaluation and whether a kill switch can make advanced AI safer. Good Morning America Sep 21, 2026 Media Mention HAI Policy Fellow Jennifer King advises caution surrounding one's person security when using AI, citing her research on AI privacy and insights on user

匹配字段正文

rts Weigh the Risks, Rules and Race Ahead Shana Lynch Sep 22, 2026 News In a new series called Prompt Response, Stanford’s Surya Ganguli, Diyi Yang and Rob Reich examined emergent agent behavior, recursive self-improvement, independent evaluation and whether a kill switch can make advanced AI safer. In a new series called Prompt Response, Stanford’s Surya Ganguli, Diyi Yang and Rob Reich examined emergent agent behavior, recursive self-improvement, independent evaluation and whether a kill switc

匹配字段正文

ew Stanford research finds they often don't measure what they claim to. In a new series called Prompt Response, Stanford’s Surya Ganguli, Diyi Yang and Rob Reich examined emergent agent behavior, recursive self-improvement, independent evaluation and whether a kill switch can make advanced AI safer. Policymakers, academics, healthcare providers, AI developers, and patient advocates convened by Stanford HAI identify critical gaps in how we regulate AI tools used for therapy and emotional support.

匹配字段正文

nstrations sometimes showcase household chores, for example, while robot makers may prioritize factory work useful to their own operations, much as AI companies prioritized coding agents.⁴⁶ Overall, robots would need to sustain record rates of price declines and quality improvements over the coming decades to enable rapid physical automation. Even so, job impacts are less certain with these cost projections, since they set aside forces like preferences and regulation as well as feedback effects

匹配字段正文

struction. A GPT-5.5 captioner writes captions for a fixed 20-video development set with 273 human-verified assertions. Gemini-3.1-Pro acts as a physics-aware critic. An evolution agent reads the scores and claim-level failures, then rewrites the prompt. The critic scores 2 dimensions: Every revised prompt is validated on PhysCapBench , a new benchmark of 246 videos and 3,794 human-verified assertions. Caption F1 rose from 78.64 at iteration 1 to 87.82 at iteration 9. The path was not smooth. It

匹配字段正文

reats an automated reply as blanket permission GPT-6 Astra frequently asked for permission before carrying out unauthorized actions. Because the evaluation simulated an autonomous agent, it always received the same automated reply. "Please proceed to the next step using your best judgement." The reply does tell the model to continue, but GPT-6 Astra treated it as blanket approval for actions clearly outside the evaluation scope, even for steps it hadn't even asked about. In some cases, the model

匹配字段正文

Google Cloud AI Research , with UNC-Chapel Hill, Stanford and Washington University in St. Louis, has released RRSI (Regularized Recursive Self-Improvement) . It lets an LLM agent rewrite its own harness: prompts, tools, memory, control flow and sub-agents. Model weights never change. RRSI constrains the improvement loop itself, so gains hold on benchmarks the agent never optimized against. Deployable? Yes, as a research framework. The code is Apache 2.0 , needs Python 3.10+, and accepts any Lit

匹配字段正文

生时代软件工程的核心管控框架。 2. 产业格局：全球竞逐转向底层，国内开源底座突破 随着全球驾驭工程应用层工具能力趋同，差异化空间收窄，竞争焦点转向底层框架。海外，Anthropic等率先完成工业级验证并形成商业产品矩阵，OpenAI于2026年8月开源Codex Harness，谷歌、微软等厂商同步推进，形成多家并行的竞争态势。国内，产品化与开源并进，华为AgentArts长程任务完成率80%，DeepSeek开源Harness框架24小时获约5.5万星，填补国内驾驭框架开源空白。 3. 技术体系：驾驭工程六大能力协同，能力逐级递进 报告提出连接、编排、循环、效能、安全、评估优化六大驾驭工程核心能力，分层协同贯通智能体与外部系统语义互通、长任务过程控制、自主循环迭代、安全审计、效能优化与评估迭代的技术全链路。在此基础上构建L1-L5能力成熟度模型，沿工具接入、运行底座、治理闭环、协同自治到规模生产逐级演进，为企业提供可评可验的建设路径。 4. 价值赋能：研发全流程可信可控，支撑多行业落地 驾驭工程将安全合规、架构标准、质量指标等要求原子化嵌入需求、编码、测试、运维等研发全流程契约，成

匹配字段正文

ollowing the release of MiMo-V2.6, tool-call repetition emerged as one of the most noticeable issues affecting the user experience. In MiMo Desktop, MiMo Code, OpenCode, and other agentic settings, the model would sometimes issue the same or highly similar tool calls repeatedly, consuming substantial time and context without making meaningful progress. Our internal evaluations confirmed this pattern: the response-level repetition rate exceeded 0.05%. The table below breaks down the rates for MiM

匹配字段正文

ing, but there are others. Kahlow mentioned robotics — for example using a simulated environment to test how a robot works. Another, more intriguing, use case is to use it to test agents at scale . “Having thousands of simulated environments is much less challenging if you have a suitable model like GWM Worlds,” Kahlow said. But how does an agent know what’s changed in the world — is there a structured state that it can read, or is it just the generated video and audio that it’s consuming and un

匹配字段正文

Summary To see what works and what breaks when agents are sent into a market, we made a miniature market of Claudes—a more controlled sequel to Project Deal , our first experiment with agents interacting in a marketplace on people's behalf. Anthropic employees across six offices brought in a book they wanted to give away. Each participant had a short chat with Claude about what they like to read, and sent a Claude-powered agent onto an open trading floor to pitch, haggle, and strike deals with o

匹配字段正文

PLUS: Your go-to guy lives in the browser Good morning, AI enthusiasts, and welcome to our 7,336 new readers. Claude spent less than 24 hours sifting through DNA data. 950 agents and 210M tokens later, it had a new mystery for the humans. Anthropic said its AI biology lab just made its first discovery, identifying an unfamiliar system inside viruses that could offer new gene-editing clues. What does it actually do? Even Anthropic doesn't know yet. In today’s AI rundown: Claude spots a CRISPR-lik

匹配字段正文

her general AI models can systematize and accelerate such discoveries . We believe that this acceleration will come from establishing a new way of doing biology research, in which agents collaborate with humans in every step of the process. Developing this new way of working required that we build our own lab and a single team working on everything from training Claude in biology to running experiments in the lab. Today, we’re sharing early results from one of our first research programs, in whi

匹配字段正文

pected to grow over time, though there are reasons to be skeptical of the magnitude. We show preliminary results on a prototype evaluation that tests monitors' ability to catch AI agents doing side tasks, and AI agents' ability to bypass this monitoring. We build on our time-horizon work and analyze 9 benchmarks for scientific reasoning, math, robotics, computer use, and self-driving in terms of time-horizon trends; we observe generally similar rates of improvement to the 7-month doubling time i

匹配字段正文

–antigen binding affinity prediction and candidate recommendation; Context-aware multi-property antibody predictor: A novel framework integrating text and protein language models; Agent-guided de novo design of nanobody binders against a novel cancer target Amazon Bio Discovery has developed three AI approaches to accelerating antibody drug design: MochiBind for ranking binding candidates by sequence alone, CA-MAP for predicting antibody developability properties while accounting for lab-specifi

匹配字段正文

Join us for our inaugural conference, Forge 2026 PUBLISHED 9/21/2026 How much better could a coding agent perform if it used the best model for each task? The best single model, GPT-6 Astra, gets 74.1% of DeepSWE tasks at $6.52 each. Pick the right model for each task and the same eighteen models get 97.6% at $1.88. 23 points better, at under a third of the cost. That number comes from hindsight. We ran all eighteen models on every task first and picked the winner for each one. What it measures

匹配字段正文

New Model Architecture Implications for TAM of DRAM/NVMe, DeepSeek V4.1 Flash, AgentX, InferenceX, NVMe experiments Engram extends standard token embeddings with learned multi-token lookups. Recurring local patterns retrieve vectors directly, reducing the need to reconstruct them through attention and feed-forward layers. With Engram model architecture optimization, it allows for lower HBM capacity to be needed for models at the same quality. This does not mean there won’t be an insane demand fo

匹配字段正文

the memory utilization of these models, making it possible to predict biomolecular systems of unprecedented sizes. By combining these results with simplifications to our previous agentic protein design approach, we show that Claude can achieve comparable in silico performance to the results we previously reported using two orders of magnitude fewer GPU hours. Beyond protein design, these specialized biological models are widely used by molecular biologists, including for drug discovery and deve

匹配字段正文